Method and apparatus for representing multidimensional data
The present invention relates to methods for representing multidimensional data. The methods of the present invention are well suited but not limited to the representation of multidimensional data in such a way as to enable the comparison and differentiation of data sets. For example, the invention may be applied to the representation of flow cytometric data. The invention further relates to a program storage device having instructions for controlling a computer system to perform the methods, and to a program storage device containing data structures used in the practice of the methods.
1. A computer-implemented method of generating a vector of values comprising a multi-resolution characterization of flow cytometry data, the method comprising:
identifying, at a server comprising a memory and a processor, flow data comprising events that describe a quantity of at least a first antibody in an individual cell, wherein the quantity is identified using flow cytometry;
generating a vector of values that characterizes the flow data responsive to iteratively identifying a series of hyperplanes used to segregate the flow data into subsets, wherein at each iteration a hyperplane is independently identified based on a subset of the flow data;
identifying the hyperplane responsive to determining a direction of the maximum variance of the subset of the flow data;
generating a rotated subset of data responsive to rotating the co-ordinates of the subset of the flow data in the direction of the hyperplane;
determining a first value of the rotated subset of the flow data;
splitting the rotated subset of the flow data into two subsets of data according to the first value;
storing the vector of values in the memory; and
uniquely identifying the flow data based on the vector of values.
2. The method of claim 1 , wherein generating the vector of values that characterizes the flow data comprises storing data that represents the hyperplane in the vector of values.
3. The method of claim 1 , wherein generating the vector of values that characterizes the flow data comprises storing the first value in the vector of values.
4. The method of claim 1 , further comprising:
aggregating the flow data associated with a plurality of samples, wherein the plurality of samples are associated with a class; and
generating a template vector of values based on the aggregate flow data according to the method of claim 1 , wherein the template vector of values represents the class.
5. The method of claim 4 , further comprising:
generating a first vector of values based on the flow data associated with a sample; and
generating a fingerprint for the first vector of values based on the template vector of values.
6. The method of claim 5 , wherein generating a fingerprint for the first vector of values based on the template vector of values comprises:
generating a score that represents a value in the vector of values relative to a probability density associated with the template vector of values.
7. The method of claim 5 , wherein the fingerprint is a binary fingerprint.
8. The method of claim 5 , wherein the fingerprint is a categorical fingerprint.
9. The method of claim 4 , further comprising:
generating a first template vector, wherein the first template vector represents a first class;
generating a second template vector, wherein the second template vector represents a second class; and
determining, based on the first template vector and the second template vector, a set of values that uniquely identify the first class and a set of values that uniquely identify the second class.
10. A computer-implemented method of classifying flow cytometry data, the method comprising:
identifying, at a server comprising a memory and a processor, flow data comprising events that describe a quantity of at least a first antibody in an individual cell, wherein the quantity is identified using flow cytometry;
generating a vector of values that characterizes the flow data responsive to iteratively identifying a series of hyperplanes used to segregate the flow data into subsets, wherein at each iteration a hyperplane is identified based on the direction of the maximum variance of a subset of the flow data;
generating a rotated subset of data responsive to rotating the co-ordinates of the subset of the flow data in the direction of the hyperplane;
determining a first value of the rotated subset of the flow data;
splitting the rotated subset of the flow data into two subsets of data according to the first value;
storing the vector of values in the memory; and
determining whether the flow data belongs to a class based on the vector of values and a template vector that represents the class.
11. The method of claim 10 , wherein the template vector is generated responsive to:
generating a plurality of vectors of values based on the flow data derived from samples associated with the class; and
generating the template vector that represents the class responsive to combining the plurality of vectors.
12. The method of claim 10 , wherein the template vector represents a class of reproducible data and the step of determining whether the flow data belongs to the class comprises determining a quality of the flow data.
13. The method of claim 10 , wherein the template vector represents a class of samples associated with a disease state.
14. The method of claim 13 , wherein the step of determining whether the flow data belongs to the class comprises determining a diagnosis associated with an individual.
15. The method of claim 13 , wherein the step of determining whether the flow data belongs to the class comprises determining a prognosis associated with an individual.
16. The method of claim 13 , wherein the step of determining whether the flow data belongs to the class comprises determining a stage of disease progression associated with an individual.
17. A computer-implemented method of generating a vector of values comprising a multi-resolution characterization of flow cytometry data, the method comprising:
identifying, at a server comprising a memory and a processor, flow data comprising events that describe a quantity of at least a first antibody in an individual cell, wherein the quantity is identified using flow cytometry;
generating a vector of values that characterizes the flow data responsive to iteratively identifying a series of hyperplanes used to segregate the flow data into finer-resolution subsets, wherein at each iteration a hyperplane is independently identified based on a finer-resolution subset of the flow data;
identifying the hyperplane responsive to determining a direction of the maximum variance of the subset of the flow data;
generating a rotated subset of data responsive to rotating the co-ordinates of the subset of the flow data in the direction of the hyperplane;
determining a first median value of the rotated subset of the flow data;
splitting the rotated subset of the flow data into two subsets of data according to the first median value;
storing the vector of values in the memory; and
uniquely identifying the flow data based on the vector of values.